Cheap DIY Soldermask - exposing with a UV laser engraver.

[Fflint]

Is anyone using their UV laser engraver for exposing/removal of the Chinese soldermask? (Usually "Mechanic" brand)?

I had it work pretty well on small testing areas and before I try to do an entire pcb I wonder if anyone can share any tips they discovered.

For those that don't know why I'm doing it, or how it works here is the background. I recently had a need to make some DIY pcbs with very small features (0.25mm pads, same clearance - 7x7 mm 44pin qfn chips etc) and soldermask. I managed to do it with a cheap cnc router and 0.1mm 15deg engraving bits. However I have no spring loaded holder so I had to very precisely adjust the Z height for Soldermask removal (I found 0.04 above the copper surface works very well - also using a height map is a must). I could make one as I have access to a lathe, but I thought it wouldn't be precise enough anyway(being spring loaded).

It took 5 tries to get the Z height right and the whole process depends on the pcb not lifting even 0.01mm so I was interested in finding something better.

I have a "3.5W" 405nm laser engraver for the same router so I decided to give it a try. Unfortunately removal of the soldermask with a laser is difficult, firstly because of the horrible toxic fumes (instant headache despite ventilation), second I only tried few power settings and I either burned the pcb along with the soldermask or not removed it fully.

Then I tried exposing it. This worked pretty well with about 10% power and 300mm feed speed (grbl in laser mode so it adjusts laser power based on movement speed). However what worked even better was lifting the power to 15% with the same speed. This resulted in burning of the topmost layer of the Soldermask, but underneath there is a thin very well exposed layer.

The problem with the lower setting was that resulting exposed areas were still quite soft and wiping with solvent it was difficult not to damage them. Also, it is very difficult to spread the mask thin enough for qfn chips not to lift too high above pads. Using the higher power it solves both problems. There are still some fumes, but nowhere near as strong as before.

For those that try doing it themselves I say this. Set too high power first (like 30% for a 3.5W laser) and do 5x5 mm squares bumping the power down by 5%. When the power is too high the Soldermask will be burned black and there will be bleed through to the sides. Then as you drop power down it becomes more of a grey color rather than black. At this stage there is bleed through to sides, and one can wipe the burned layer with ISO alcohol to reveal perfectly exposed mask below.

So I decided to come here and let those that don't know that it is indeed doable, and ask if there is anyone already using it hoping someone can share more tips.

[tomgat]

I have tried this for burning off the solder mask pads on prototype boards as a time savings to photo setting, but gave up due to it being too hard to align the board properly.  I have a K40.  The biggest issue with using these lasers for PCBs is alignment. 

If I ever get a little extra time, I would like to create a microscope crosshair alignment camera system for my K40 similar to what the LPKF milling machines have.  Its such a simple thing to build (LPKF milling machines have a cheap microsoft based one with an attached lense).  Im betting someone has already done this, but to be honest I haven't looked very hard.

[IconicPCB]

I use liquid photo imageable solder mask
'Apply it with a squeegee and silk screen.
Cure it as proscribed expose it using a 150mW laser diode and then develop in calcium carbonate solution.

Results are suitable for prototype level work.
The limitation is essentially laser spot size.

0,2mm seems to be about right.

[aix]

I use liquid photo imageable solder mask
'Apply it with a squeegee and silk screen.
Cure it as proscribed expose it using a 150mW laser diode and then develop in calcium carbonate solution.

I'd love to learn more about this process.  Would you mind sharing which LPI mask product you use and, if you can, post some photos of the results?

[Fflint]

I have tried this for burning off the solder mask pads on prototype boards as a time savings to photo setting, but gave up due to it being too hard to align the board properly.  I have a K40.  The biggest issue with using these lasers for PCBs is alignment. 

If I ever get a little extra time, I would like to create a microscope crosshair alignment camera system for my K40 similar to what the LPKF milling machines have.  Its such a simple thing to build (LPKF milling machines have a cheap microsoft based one with an attached lense).  Im betting someone has already done this, but to be honest I haven't looked very hard.

Interesting. I did try burning the mask off before I settled on the current method. I couldn't stand the smell and also the power level required to burn off Soldermask would delaminate the trace underneath.

However, I recently found running two exposures is even better than one and the thinner one can get the soldermask paste to start with the better. I also found running at 60 (out of a thousand, so 6%) power of a 3.5W laser and two passes works great.

I tend to put blobs of the paste squeezing them directly out of a syringe on the board. I then heat it a bit with a heat gun to make it more liquid and then I do multiple passes with an old credit card to spread it. When the color becomes light I know it is thin enough.

I have no problem with alignment, because the board is already positioned after isolation milling of traces. I never remove it from the bed (I have alignment pins for double sided boards). My entire process for single sided boards is:

(pins are not needed for single sided boards, but I still use them in case I need to realign).
I use a cheapo chinese 3018 mini router, but I bought a couple of collet chucks. I selected the best one and I aligned it making permanent marker marks on the matter tor shaft and the chuck so runout cancells itself mostly. This way I get a runout of about 0.05mm (a bit over 1 thou) which is enough to isolation mill 0.25 mm gaps between traces.

1 - find home position with an alignment pin (my cnc has no limit switches) by using the electrical probing in x an y axis.
2 - remove alignment pins
3-  put a large piece of pcb on the router bed, put 1mm drill in the collet. Run the program to drill alignment holes.
4- take the board off, enlarge the holes to 3mm using cordless drill, put the pins in, apply two sided tape to the back of the board, put it on the bed with the pins, load a 15 degree V bit in the collet.
5- Do Z probe. Run isolation milling program, use 0000 steel wool to deburr the pcb in the end. Use a vacuum cleaner and a brillo pad with alcohol to clean it.
6 - Spread the paste with a CC heating with a heat gun if necessary, swap the router head to a laser. I have a quick head change system of my design inspired by the mutant from wham bam - mine is a lot better of course :-)
7 - run the laser exposure of the Soldermask program, clean the remaining unexposed paste with "pure cotton" tissues and alcohol (there are better solvents, but this works fine too). Swap the laser back to a milling head, put a 0.9mm carbide drill in.
8 - run a drilling program, then swap for a 2mm end mill.
9 - run a pcb shape milling program.

important stuff including speeds and feeds
Isolation milling of traces:
Carbide V bit 15deg 0.1mm end. 120mm/min xy feed, 60mm/min z feed, 400mm/min rapids, two passes 0.04mm deep each. I set the tool width to 0.2mm in the software. In reality it is 0.2~0.25 due to runout.
PCB milling:
2mm carbide end mill, same speeds and feeds, 0.1mm per pass deph, full slotting.
Drilling:
Same speeds and feeds, no pecking.
Laser exposure :
Same speeds and feeds, 6% power, two passes, grbl in M4 mode (important!)

I use altium to create gerber files and nc for drilling, then I use flatcam to generate all of my g code (I have it automated with a script). Then I use Candle for Z-probe and to adapt the gcode with probe results and to run the gcode.

I use camotics to emulate my gcode before cutting as sanity check.

I'm very happy with this process except the spreading of the soldermask with a CC. It requires a level of skill and I forget how to do it if I don't do it for few weeks wasting paste for my first try. I would prefer to have something that can do that well with a single stroke. I'm currently looking for a cheap rubber roller to try.

I use liquid photo imageable solder mask
'Apply it with a squeegee and silk screen.
Cure it as proscribed expose it using a 150mW laser diode and then develop in calcium carbonate solution.

Results are suitable for prototype level work.
The limitation is essentially laser spot size.

0,2mm seems to be about right.

I too would very much like to hear more about this product. The only two types of Soldermask I saw was dry film and the one I'm using (mechanic, UV curable, no specific developer - just solvent).